Early Reader

Most preschool reading apps compete on the same currencies: coins, stars, unlockable characters, daily streaks. Early Reader, listed on the App Store as Early Reader Preschool, refuses all of them. Published by Instalabs Limited Liability Company, it is a free, open-source phonics course for young children. It is built to be run by a parent sitting next to the child, not handed over as a distraction.

The teaching method comes from DISTAR, the direct-instruction approach developed by Siegfried Engelmann. That means explicit letter-sound teaching, carefully sequenced lessons and immediate error correction. It is a method built around an adult who delivers instruction and catches mistakes as they happen. Early Reader adds the SM-2 spaced repetition algorithm, the same scheduler behind Anki, so cards a child struggles with come back sooner and mastered cards come back less often.

The content is a finite curriculum that runs from letter sounds through digraphs and decodable words to full sentences, and every card has its own illustration. A bundled audio library supplies the spoken layer. Everything ships in a sizable download, after which the app works offline, asks for no account and reports collecting no data. Progress lives on the device.

The positioning is unusual. The source code is public on GitHub under AGPLv3, and Instalabs publishes the official store builds. That places Early Reader somewhere between a commercial children’s app and a community software project. It also makes the audience clear: parents who care more about method and privacy than about spectacle.

Evidence of how it lands with families is thin. The US App Store shows a 5.0 average from a single rating. That is too little to establish how it works for families, and no independent professional coverage exists yet.

• DISTAR-based direct instruction: Explicit letter-sound teaching, a fixed lesson sequence and immediate error correction, following Engelmann’s method.
• SM-2 spaced repetition: The algorithm used by Anki brings cards back at intervals set by how the child performs.
• Finite curriculum: Letter sounds, digraphs, decodable words and sentences.
• Illustration and audio on every card: A pre-generated illustration per card, plus bundled audio for pronunciations, prompts, encouragement and hints.
• Optional speech recognition: Real-time pronunciation checking that families can use or ignore.
• No gamification: No streaks, rewards or game layer, by design.
• Offline and account-free: Works without a connection after the sizable download, needs no login, reports collecting no data and keeps progress on the device.
• Open source: Code on GitHub under AGPLv3, with official builds published by Instalabs, LLC.

A Classroom Method, Rebuilt for the Sofa

DISTAR assumes an adult is delivering the lesson. Early Reader keeps that adult involved instead of trying to automate them away, and that is the central design decision in the app. It is also the right one for the method. Immediate error correction, the defining feature of direct instruction, requires someone to notice the error. A parent sitting beside the child can hear a muddled vowel, model the right sound and move on in seconds. The app’s job is to supply the sequence, the prompts, the pictures and the audio, so the parent never has to plan a lesson or decode a teaching manual.

The cost is just as clear: Early Reader does nothing without a grown-up. Families looking for something a child can use alone while dinner gets made have the wrong product. We read that as a deliberate filter, not a flaw, but it is a commitment parents should weigh before spending storage on it.

The Spaced-Repetition Engine

SM-2 was designed for adults reviewing their own flashcards and grading their own recall honestly. Moving it to preschoolers raises an obvious question: who does the grading? In a parent-led session, the answer is presumably the parent, with optional speech recognition as a second signal. The schedule can only be as good as those judgements. A generous parent who marks a shaky answer as solid will push that card further out than it should go, and the algorithm will trust the parent.

The pairing is sensible. DISTAR decides the order in which new material arrives, and SM-2 decides how often old material returns. That split of responsibilities fixes a long-standing weakness of linear phonics workbooks, where a sound learned early can fade unnoticed because the book has moved on. Here, a forgotten digraph comes back because the child’s own performance says it should.

Speech recognition is an ambitious feature and one we would treat cautiously. Young children’s speech, with its half-formed sounds, lisps and regional vowels, is a hard input for any recognition system. A false “wrong” on a correct answer is exactly the kind of error direct instruction exists to prevent. Making the feature optional is the right call, because it leaves the parent as the final authority.

The Curriculum: Wide at the Base, Short at the Top

The card mix reveals the teaching priorities. Decodable words dominate, which is where phonics fluency gets built: repeated blending practice with words that use only sounds the child already knows. The sound set goes beyond the alphabet’s letters, which shows the app teaches sounds rather than letter names. Digraph cards extend the code to two-letter spellings.

The ceiling is sentences. The card list contains no stories or longer passages, so Early Reader is a bridge to books, not a replacement for them. Parents should plan for the library visit that comes next. The finite card set also means the app eventually runs out of new material. After that, spaced repetition can only keep reviewing what the child has already learned.

The No-Reward Bet

Leaving out streaks and rewards is a philosophical stance, and we think it holds up. Streak mechanics bring adult engagement anxiety into a preschooler’s day. Reward loops risk making the token the goal and the reading the chore. Early Reader keeps encouragement in the audio, since praise and hints are part of the bundled clips, so a child hears approval without collecting anything.

The trade-off falls on the parent. Motivation now comes from the adult in the room and from the satisfaction of reading a word unaided. Some children will find that enough. For others, a session that ends without a star may be a harder sell, and the parent will have to supply the momentum.

Privacy and Open Source as Features

No account, offline operation, no reported data collection and on-device progress leave very little for anyone to harvest, which matters most for a children’s app. The AGPLv3 license strengthens the privacy claim: anyone can read the code, and modified versions must stay open. Inspectable source is a stronger privacy guarantee than a policy page. It also protects families against a small developer walking away, since the code and curriculum stay public even if the store build lapses.

The downside of device-only storage is that no account means no account-based sync. A child’s review schedule will not follow them to a second device.

Every card has a pre-generated illustration. For a child decoding a word they may not recognise, a picture anchors the meaning. Phonics purists will raise a fair concern, though: an image can let a child guess the word instead of sounding it out. How much that matters depends on when the picture appears relative to the word.

With no currencies, avatars or reward screens to accommodate, the interface has far fewer jobs than a typical preschool app’s. That simplicity suits a screen two people share. The App Store screenshots are the only visual record available; no independent coverage exists.

On performance, the design choices point in a sensible direction. All illustrations and audio come pre-generated and bundled, so nothing needs to stream mid-session. That removes the network dependence that often causes stalls in connected children’s apps. The local-first SQLite approach is a sensible fit for a bundled card app. App size and load times are fair questions for a large download; a flashcard app of this weight will claim real storage on older or base-model iPhones.

Early Reader is an iOS app, and its design follows from that:

• Touch on a shared screen: The parent-and-child format means two people looking at one display. A shared-screen app benefits from the largest screen in the house.
• Microphone optional: The speech recognition feature presumably relies on microphone access. Because it is optional, families who decline that permission keep the full curriculum.
• Offline by design: After the initial download, the app works on a plane, in a car or in a bedroom with weak Wi-Fi. That fits a short parent-led session at home or away from Wi-Fi.
• Price: The app is free, and the GitHub source gives technically minded parents a way to inspect what runs on their child’s device.
• No cross-device sync: With no account, progress stays on whichever iPhone or iPad the child uses.

Early Reader makes a clear argument: a proven teaching method, a sensible review algorithm and a parent beside the child beat any amount of animated reward. We think the argument is sound. DISTAR’s sequenced lessons and immediate correction suit a format where an adult is present, and SM-2 addresses the forgetting problem that linear workbooks ignore. Leaving out streaks and points is a principled choice, not an oversight. Free distribution, offline operation, no account and open AGPLv3 code make the value case straightforward.

The reservations are real, though. The app demands adult time every session. Its curriculum stops at sentences, leaving the move to books entirely to the family. Its optional speech recognition takes on one of the hardest inputs in voice technology. Above all, the evidence that it works is a single App Store rating.

Parents who already believe in explicit phonics and can protect a few minutes a day will find a well-reasoned, privacy-respecting tool here at no cost. Parents who need the app to do the teaching for them should look elsewhere.